Big Data Can't Fix Your Customer Service...

customer-service...Unless you are actually resolved to serve your customers well.

Many businesses today are excited about the possibilities of big data and analytics to enable them to be more responsive to their clients and prospects. Some of the promised benefits are:

  1. Better retention and reduced churn - The most profitable customer is one you can keep long term. Every company should be able to estimate the lifetime value of each client and compare to the cost of new customer acquisition. Nurture marketing is way less expensive than the alternative of finding replacements customers.
  2. More marketing leverage - Your happier clients are more likely to make repeat or increased purchases, and further they will tell their social circle about their experiences, conditioning a broader part of the market to be more receptive to your offers. Unhappy customers will share their negative experiences, and usually more forcefully as humans naturally react more strongly to learned "dangers." This is why net promoter scoring is so popular.
  3. Lower cost of customer support - Trying to win back a disgruntled customer can be expensive in several ways. First, it simply takes time from your agents, who could otherwise be upselling and recruiting incremental business. Second, it usually often involves giving away some value to make amends, whether in discounts or refunds. Third, it's damaging to your own customer service team morale, leading to lower quality employee work and higher staff turnover.

The good news is there are now many big data techniques and tools available to help, including:

  1. Customer 360' profiling and segmentation - By combining multiple information sources, such as service calls with billing history with client profile information, you can have a much more informed view of how to handle different situations. Nothing annoys a client more than realizing each representative they speak with has no idea who they are or what their recent issues have been. Conversely, you can define guidelines for how to satisfy high value clients with customized service and offers.
  2. Real-time, emotional, and social media analytics - Reacting in the moment is critical to good service. Knowing why they are calling is as important as understanding how they feel about the situation. Empathy matters tremendously. If you can see what they are saying, and link this up across multiple communications media such as phone, e-mail, and social networks, you are in much better shape to respond appropriately and in time to defuse an issue.
  3. Predictive and advanced analytics - Defining common conditions and cohorts will help to predict the likely outcomes, and especially the implications of those results. Will this customer leave you and take their business elsewhere? Will your reputation be damaged if they tell others? What is the lost revenue, immediately and in the future? Being able to answer these questions will more accurately guide your response.

Let's take a case study of failure from my own experience here. I recently had a car stolen, my dear 1965 Ford Mustang. Luckily, the police recovered it two days later and the damage was minimal: only $250 in parts and I was able to repair it myself in a few hours. I was upset at the thieves, but figure it was just some dumb kids joy-riding.

What really infuriated me was the series of blunders by my long-time insurance company, who won't be named, but let's just say they operate in "all states" of the country. When I called to report the theft, we discovered that their agent had configured not one but TWO different insurance policies on this car and had been double-billing me for three years, at an estimated cost to me of $1,000. I can’t be sure because I can’t find the details of how much was being charged per car. Maybe this is not a lot of money, but if you’re keeping score, I’ve lost FOUR times as much money to incompetence of the agent as to the car thieves. It shouldn’t be hard to notice that they’ve set up redundant policies if they have a basic profile of my accounts.

On the day my car was found, I went to recover it from the police in a bad neighborhood. I called the insurance company for help on the scene and was told by their representative that they cannot send a tow truck on Sundays, they cannot recommend a local garage to fix issues, and they also cannot connect me with a claims adjuster on Sundays. Then they invited me to stay on the line for a survey and wanted to take some time to update my contact address. While I’m on the side of the road in a rough hood with the police waiting on me. This shows not just a failure to assist as expected, but a profound lack of real-time situational awareness.

This prompted me to complain via Twitter, and this is where the company's response went from bad to worse. Their agent charged with social media response started by asking for me to tweet my full name, phone number, and claim number. Um, what? Anyone with a rudimentary understanding of data privacy and governance would see why this isn’t a good idea. Of course when I proposed he instead tweet his full name, phone number, and employee ID, then the company suddenly became concerned about privacy.


Interesting double-standard, but even more frustrating that he couldn’t match the “Nik Rouda” on Twitter to the “Nikki Rouda” who had called in that morning with a claim. Eventually, he asked me about another issue from last fall where they had denied responsibility for a different claim. Bringing up past service failures doesn’t help improve your customer’s mood when they really want to solve the problem at hand.

In the aftermath, I had asked this sorry-excuse-for-an-insurance-company to refund the extra billing for the unneeded policy, cancel my policy, and end the claim since I had already fixed the damages myself. Then the company sent me another bill for the upcoming renewal period. AGAIN including charges for a redundant policy I most clearly did not want, now or anytime in the past. The reason given? They didn’t want to disrupt the claim for which they had foisted responsibility and which I had already told them to drop much earlier. This was now in the realm of farce.

Long story short (too late!) they have refused to refund the double billing, did not help at all when I needed them, and, through a comedy of avoidable errors, has exacerbated the situation and created enormous ill-will. So now I’ll tell this story to my 1,300+ colleagues on LinkedIn, my 800+ followers on Twitter, my 180+ close friends on Facebook, and anyone who ever asks me about auto insurance.

Some pretty simple solutions involving big data and analytics could have avoided almost all of these problems. For final context, I’ve spent somewhere over $100,000 (!) with this company over the last dozen years in homeowners, earthquake, umbrella, and multiple auto policies. I should be considered a pretty darn good customer, yet I now feel a tremendous betrayal of the trust I placed in them. All the technology in the world won’t help if the business isn’t really interested in serving the needs of their customers.

big data analysis

Topics: Data Platforms, Analytics, & AI